Site Search Engine for 1,000 page website
- by Ian
I manage a website with about 1,000 articles that need to be searchable by my members.
The site search engines I've tried all had their own problems:
Fluid Dynamics Search Engine
Since it's written in perl, it was a bit hacky to integrate with my PHP-based CMS. I basically had to file_get_contents the search results page. However, FDSE had the best search results.
Google CSE
Ugh, the search results SUCK. It can't find documents even using unique strings. I'm so surprised that a Google search product is this bad. Nor can I get any answers on their 'help' forums, and I am a paying user. Boo, Google. Boo.
Sphider
Again, bad search results. Unable to locate some phrases used in link text. Better results than Google CSE though. Shame on Google that a free PHP script has better search results than their paid application.
IndexTank
This one looked really promising. I got all set up with their PHP API client. But it would only randomly add articles that I submitted. Out of 700+ articles I pushed to the index through their API, only 8 made it in. Unable to find any help on this subject.
Update for IndexTank -- Got the above issue fixed, so this looks most promising so far.
The site itself runs on php/mysql and FreeBSD, though this shouldn't matter for a web crawling indexer.
I've looked at Lucene, but I don't know anything about Java or installing Java programs on my web server.
I also do not have root access on my web server, if this would be required for installation.
I really don't need a lot of fancy features. It just needs to be able to crawl my web site and return great (even decent!) search results. I don't need any crazy search operators. It doesn't need to index off my primary domain. It just needs to work!
Thanks, Hive Mind!
